Les applications récentes telles que les bases de données forestières impliquent des observations de données spatiales associées à l'observation de nombreuses covariables spatiales. Nous considérons dans cette thèse le problème de l'estimation d'une forme paramétrique de la fonction d'intensité dans un tel contexte. Cette thèse développe les procédures de sélection des variables et donne des garanties quant à leur validité. En particulier, nous proposons deux approches différentes pour la sélection de variables : les méthodes de type lasso et les procédures de type Sélecteur de Dantzig. Pour les méthodes envisageant les techniques de type lasso, nous dérivons les propriétés asymptotiques des estimations obtenues par les fontions d'estimation dérivées par les vraisemblances de la Poisson et de la régression logistique pénalisées par une grande classe de pénalités. Nous prouvons que les estimations obtenues par de ces procédures satisfont la consistance, sparsité et la normalité asymptotique. Pour la partie sélecteur de Dantzig, nous développons une version modifiée du sélecteur de Dantzig, que nous appelons le sélecteur Dantzig linéaire adaptatif (ALDS), pour obtenir les estimations d'intensité. Plus précisément, les estimations ALDS sont définies comme la solution à un problème d'optimisation qui minimise la somme des coefficients des estimations soumises à une approximation linéaire du vecteur score comme une contrainte. Nous constatons que les estimations obtenues par de ces méthodes ont des propriétés asymptotiques semblables à celles proposées précédemment à l'aide de méthode régularisation du lasso adaptatif. Nous étudions les aspects computationnels des méthodes développées en utilisant les procédures de type lasso et de type Sélector Dantzig. Nous établissons des liens entre l'estimation de l'intensité des processus ponctuels spatiaux et les modèles linéaires généralisés (GLM), donc nous n'avons qu'à traiter les procédures de la sélection des variables pour les GLM. Ainsi, des procédures de calcul plus faciles sont implémentées et un algorithme informatique rapide est proposé. Des études de simulation sont menées pour évaluer les performances des échantillons finis des estimations de chacune des deux approches proposées. Enfin, nos méthodes sont appliquées pour modéliser les emplacements spatiaux, une espèce d'arbre dans la forêt observée avec un grand nombre de facteurs environnementaux. / Recent applications such as forestry datasets involve the observations of spatial point pattern data combined with the observation of many spatial covariates. We consider in this thesis the problem of estimating a parametric form of the intensity function in such a context. This thesis develops feature selection procedures and gives some guarantees on their validity. In particular, we propose two different feature selection approaches: the lasso-type methods and the Dantzig selector-type procedures. For the methods considering lasso-type techniques, we derive asymptotic properties of the estimates obtained from estimating functions derived from Poisson and logistic regression likelihoods penalized by a large class of penalties. We prove that the estimates obtained from such procedures satisfy consistency, sparsity, and asymptotic normality. For the Dantzig selector part, we develop a modified version of the Dantzig selector, which we call the adaptive linearized Dantzig selector (ALDS), to obtain the intensity estimates. More precisely, the ALDS estimates are defined as the solution to an optimization problem which minimizes the sum of coefficients of the estimates subject to linear approximation of the score vector as a constraint. We find that the estimates obtained from such methods have asymptotic properties similar to the ones proposed previously using an adaptive lasso regularization term. We investigate the computational aspects of the methods developped using either lasso-type procedures or the Dantzig selector-type approaches. We make links between spatial point processes intensity estimation and generalized linear models (GLMs), so we only have to deal with feature selection procedures for GLMs. Thus, easier computational procedures are implemented and computationally fast algorithm are proposed. Simulation experiments are conducted to highlight the finite sample performances of the estimates from each of two proposed approaches. Finally, our methods are applied to model the spatial locations a species of tree in the forest observed with a large number of environmental factors.

Les crues – telles que les laves torrentielles – engendrées dans les torrents lors de fortes précipitations peuvent mobiliser de grande quantité de sédiments. Lorsqu'elles atteignent les zones urbanisées, elles peuvent mettre en dangers à la fois les personnes et les biens. Les approches visant à réduire le risque torrentiel se basent largement sur des seuils intensité-durée de pluie qui déterminent les conditions minimum de déclenchement d’une lave torrentielle. Pourtant, ces seuils sont sujets à une forte variabilité liée, non seulement aux différences inter-sites, mais aussi à la méthode appliquée lors de leur établissement. De plus, ils peuvent entraîner des fausses prédictions, l’intensité et la durée de l’épisode de pluie n’étant pas les seules variables explicatives. Ce travail de thèse vise (i) à fournir un cadre méthodologique rigoureux pour l’établissement des seuils de pluie afin de limiter les sources de variabilité, et (ii) à améliorer leurs performances en considérant à la fois les facteurs de déclenchement et de prédisposition. Il s’appuie sur les données d’un observatoire des crues torrentielles, mis en place dans les Alpes françaises en 2011 sur les torrents très actifs du Manival et du Réal. Dans un premier temps, les images et mesures hautes-fréquences collectées entre 2011 et 2016 ont été analysées afin de détecter et de caractériser les crues torrentielles. Pour appréhender la diversité des écoulements observés, une classification phénoménologique a été proposée. Dans un second temps, la condition minimum intensité-durée de pluie requise pour déclencher une lave torrentielle a été établie. La sensibilité du seuil à la définition d’un épisode de pluie a été évaluée. Dans un troisième temps, un modèle de régression logistique a été implémenté pour discriminer les épisodes de pluies critiques qui n’ont pas engendré de lave torrentielle. Il a permis de sélectionner les variables explicatives les plus pertinentes. Finalement, des pistes de travail ont été avancées pour (i) passer de conditions critiques établies à une échelle locale vers une échelle régionale, en perspective d’une application au sein d’un système d’alerte dédié aux risques hydrométéorologiques, et (ii) passer des conditions de déclenchement d’une lave torrentielle dans la zone de production sédimentaire aux conditions de propagation jusqu'aux zones à enjeux. / Flows – such as debris flows – caused by heavy rainfalls in torrents can mobilise a huge amount of sediments. When they reach the urbanised areas, they may endanger the people’s safety or cause damages. Approaches aimed at mitigating torrential risk widely rely on rainfall intensity-duration thresholds which determine the minimum debris-flow triggering conditions. However, these thresholds suffer from a high variability related not only to inter-site differences but also to the method applied to design them. In addition, they are likely to cause false prediction because the intensity and the duration of the rainfall event are not the only explanatory variables. This PhD research work aim (i) to provide a rigorous methodological framework for designing rainfall threshold in order to limit the variability sources, and (ii) to improve their performances by including both the triggering and the predisposing factors. It is supported by field observations stemming from high-frequency monitoring stations installed since 2011 on two very active debris flow-prone torrents in the French Alps: the Manival and the Réal. First, the images and data gathered between 2011 and 2016 were analysed in order to detect and characterise the sediment laden-flows. To deal with the variety of recorded flows, a phenomenological classification was performed. Second, the minimum intensity-duration threshold for debris-flow triggering was assessed. The threshold sensitivity to the rainfall event definition was estimated. Third, a logistic regression model was used to discriminate the critical rainfall events which do not lead to a debris flow. It makes it possible to select the most relevant explanatory variables. At last, several avenues of work were proposed (i) to move the knowledge of debris-flow initiation conditions from a local to a regional level, with a view to application in a warning system dedicated to hydrometeorological risks, and (ii) to improve the ability to predict, not the debris-flow triggering in the production zone, but the debris-flow propagation up to the area concerned.

No description available.

No description available.

Thesis deals with complex process of creation of new bankruptcy model for predicting business failure, while this process involves selection of quality sample, verification of classification accuracy of already existing bankruptcy models, profile analysis and finally the derivation of specific equation of bankruptcy model. The derivation is performed by using two selected statistical methods, discriminant analysis and logistic regression. Two bankruptcy models Bürger's index DA12 and Bürger's index LR12 were derived by using the mentioned statistical methods. The new models distinct advantage is, unlike already existing and renowned bankruptcy models, that they are focused on classification of micro and small enterprises in terms of Czech Republic, while classification accuracy one year before failure is by individual models 74.8 % and 81.87 %. Derived models have clear interpretation (no grey zone) and easy calculation, which brings a possibility for micro and small entrepreneurs to check their business partners in terms of failure prediction.

Submitted by Marcos Vinícius Alvarenga Ramos da Silva (marcosvarsilva@gmail.com) on 2015-08-19T01:38:31Z No. of bitstreams: 1 Tese_V11.pdf: 1831740 bytes, checksum: ea23632eaf2595347fb5bf42e94cf418 (MD5) / Approved for entry into archive by Renata de Souza Nascimento (renata.souza@fgv.br) on 2015-08-19T15:50:35Z (GMT) No. of bitstreams: 1 Tese_V11.pdf: 1831740 bytes, checksum: ea23632eaf2595347fb5bf42e94cf418 (MD5) / Made available in DSpace on 2015-08-19T16:21:04Z (GMT). No. of bitstreams: 1 Tese_V11.pdf: 1831740 bytes, checksum: ea23632eaf2595347fb5bf42e94cf418 (MD5) Previous issue date: 2015-08-11 / Neste trabalho será apresentada a modelagem por regressão logística, com a finalidade de prever qual seria a inadimplência dos clientes que compõem o portfólio de uma grande instituição financeira do país. Sendo assim, será explorada a ideia de usar o conceito de provisionamento pura e simplesmente, através da estimação de uma probabilidade de default dado por um ou mais modelos estatísticos que serão construídos no decorrer do trabalho, conforme incentiva o comitê de Basileia. Um dos modelos será feito a partir de uma separação prévia de público através de clusters e a outra técnica a ser explorada será a criação de um modelo sem nenhuma separação. O objetivo será a comparação entre as duas métricas de classificação de risco e verificar os trade-off entre elas e os impactos de variáveis macroeconômicas nestes modelos. / This work presents the modeling logistic regression, in order to predict what the default of customers that make up the portfolio of a major financial institution in the country. Thus, the idea is exploited to use the concept of provisioning pure and simply, by estimating a probability of default data for one or more statistical models to be constructed during this work, as encourages Basel committee. One of the models will be done from a previous separation of the public through clusters and other technique being explored is the creation of a model with no separation. The goal will be to compare the two risk rating metrics and check the trade-off between them and the impacts of macroeconomic variables in these models.

Submitted by Guilherme da Silva Ferreira (guilhermesferreira053@gmail.com) on 2018-02-07T03:03:01Z No. of bitstreams: 1 MODELO DE PREVISÃO DE ENTRADA EM RECUPERAÇÃO JUDICIAL.pdf: 1209184 bytes, checksum: c2e3533402e1c78c0ddba7546a60c44b (MD5) / Approved for entry into archive by Thais Oliveira (thais.oliveira@fgv.br) on 2018-02-15T19:00:09Z (GMT) No. of bitstreams: 1 MODELO DE PREVISÃO DE ENTRADA EM RECUPERAÇÃO JUDICIAL.pdf: 1209184 bytes, checksum: c2e3533402e1c78c0ddba7546a60c44b (MD5) / Made available in DSpace on 2018-02-16T11:16:23Z (GMT). No. of bitstreams: 1 MODELO DE PREVISÃO DE ENTRADA EM RECUPERAÇÃO JUDICIAL.pdf: 1209184 bytes, checksum: c2e3533402e1c78c0ddba7546a60c44b (MD5) Previous issue date: 2017-12-08 / A crise observada na economia brasileira nos últimos anos chamou atenção para o direcionamento de crédito dos bancos e instituições financeiras, que reduziram o volume disponível para financiamento de curto e longo prazo liberados para as empresas. Diante deste cenário, modelos estatísticos se apresentam de suma importância com a função de ajudar na correta distribuição de créditos e redução de perdas com calotes. Neste contexto, foi desenvolvido um modelo para previsão de entrada em recuperação judicial utilizando informações contidas no balanço patrimonial das empresas e variáveis macroeconômicas através de regressão logística. Os resultados obtidos indicam que empresas com dificuldades financeiras são sensíveis à piora do cenário econômico. / The economic crisis observed in Brazil in the last years put a light in the form the banks and financial institutions direct their capital. The first reaction was reduction on loan volume and then, make better decision in the sense of for who make loan. In front of this scenario, statistic models are very important for help in the correct distribuition of credit and reduction of default. In this work, a model using logistic regression was builded to forecast the companies that have great chance to go bankrupt. The results showed that companies with financial stress are sensitive to the worsenig of the economic sceario.

Submitted by marcos ito (marcos_ito@yahoo.com.br) on 2012-01-20T18:55:07Z No. of bitstreams: 1 20120120 - MPA Marcos Kenji Ito.pdf: 931031 bytes, checksum: e54504c14f74f4f4d321823f6255eda4 (MD5) / Approved for entry into archive by Suzinei Teles Garcia Garcia (suzinei.garcia@fgv.br) on 2012-01-20T18:58:07Z (GMT) No. of bitstreams: 1 20120120 - MPA Marcos Kenji Ito.pdf: 931031 bytes, checksum: e54504c14f74f4f4d321823f6255eda4 (MD5) / Made available in DSpace on 2012-01-20T19:00:45Z (GMT). No. of bitstreams: 1 20120120 - MPA Marcos Kenji Ito.pdf: 931031 bytes, checksum: e54504c14f74f4f4d321823f6255eda4 (MD5) Previous issue date: 2011-12-20 / This research aims to identify a statistical model of propensity to revolve through store credit card, in order to maximize the credit appliance and optimize marketing efforts. The descriptive study aims to generate insights for understanding the expansion of consumer credit, in a context of lack of financing options and limitation in distribution channel. There will be a quantitative approach in the customer data base of a financial institution and through the study of socio demographic and transactional it pretends to establish a mathematical model followed by a validation of their predictive ability. / Esta pesquisa tem o objetivo de identificar as variáveis e sua influência na propensão à aquisição de crédito pessoal, propondo um modelo estatístico de propensão ao financiamento por cartões de crédito híbridos para maximização de contratação de crédito e otimização dos esforços de marketing. O estudo descritivo pode gerar insights para a compreensão da expansão do crédito ao consumo, sobretudo num contexto de escassez de opções de financiamento e limitação no canal de distribuição. Foram usados dados de uma base de clientes de uma instituição financeira com variáveis sócio demográficas e transacionai, e o modelo matemático foi seguido da validação de sua capacidade preditiva.

Nas últimas décadas inúmeros estudos foram produzidos acerca da influência do local de moradia sobre o bem-estar das pessoas, o chamado “efeito-bairro”. Este trabalho tem o objetivo de analisar os principais fatores que influenciam os jovens gaúchos de 15 a 24 anos a permanecerem apenas estudando, estudando e trabalhando, só trabalhando ou não estudando e nem trabalhando. Serão analisados fatores ligados ao local de moradia, características individuais dos jovens, tais como raça, sexo e idade, e familiares, tais como educação dos pais e renda familiar per capta e como cada uma dessas variáveis exerce influência sobre as escolhas dos jovens das aglomerações urbanas do Rio Grande do Sul. Para tanto, serão utilizados os microdados amostrais dos Censos Demográficos de 2000 e 2010 do IBGE, através dos quais serão construídas as variáveis dependente e independentes que constituirão a análise estatística via modelos de regressão logística multinomial. Os resultados das influências dos fatores ligados ao local de moradia apontaram que residir longe do centro, em 2000, dificultava o ingresso no mercado de trabalho para o jovem de Porto Alegre, enquanto em 2010 aumenta suas chances de estudar e de trabalhar e quanto mais elevado é o nível socioeconômico da vizinhança do jovem, maiores são suas chances de apenas estudar e menores suas chances de só trabalhar. O estudo também constatou que o jovem ser filho ou enteado do chefe domiciliar aumenta significativamente suas chances de estudar e reduz bastante suas chances de dedicarem-se exclusivamente ao trabalho; quanto mais elevada a idade do jovem, maiores suas chances de trabalhar e menores de só estudar, consequência da transição natural do jovem da escola para o mercado de trabalho; o jovem ser do sexo masculino aumenta suas chances de trabalhar; quanto mais elevada a renda domiciliar per capita do jovem, menores são suas chances de não estudar nem trabalhar; quanto mais anos de estudo o chefe domiciliar possuir, maiores as chances do jovem estudar; quanto mais crianças no domicílio do jovem, maiores são suas chances de não trabalhar nem estudar e quanto mais elevado o grupo de categoria sócio-ocupacional do chefe domiciliar, maiores são as chances do jovem estudar. / In recent decades numerous studies have been made about the influence of the place of residence on the well-being of people, the so-called "neighborhood effect." This work aims to analyze the main factors influencing the gauchos youth 15 to 24 years to stay just studying, studying and working, just working or no studying and neither working. We will analyze factors related to place of residence, individual characteristics of young people, such as race, gender and age, and family, such as parental education and family income per capita and how each of these variables influences the choices of young agglomerations urban of Rio Grande do Sul. Therefore, the sample microdata from Demographic Censuses of 2000 and 2010 IBGE will be used, through which the dependent and independent variables that constitute the statistical analysis via multinomial logistic regression models will be built. The results of the influence of factors related to place of residence indicated that reside far from the center, in 2000, made it difficult to enter the labor market for young Porto Alegre, while in 2010 increases your chances of study and work and the higher It is the socioeconomic status of the neighborhood of the young, the greater your chances of just studying and lower your chances of just work. The study also found that young to be a child or stepchild of the household head significantly increases your chances of study and greatly reduces your chances to devote themselves exclusively to the work; the higher the age of the young, the greater your chances of work and under only studying result of the natural transition of the young from school to the labor market; the young being male increases your chances of working; the higher the household income per capita of the young, the lower your chances of not study or work; the more years of schooling the household head has, the more likely the young study; the more children in the household of the young, the greater your chances of not working or studying and the higher the socio-occupational category group head home, the greater the chances of the young study.

Today biodiversity is rapidly decreasing and an increasing number of threatened species live in modified and human dominated landscapes. Therefore it is essential to learn more about how species cope with the changes of their habitat. The focus of this study lies on a primate species, the critically endangered Javan slow loris (Nycticebus javanicus), endemic to the densely populated island of Java, Indonesia.  In cooperation with the Little fire face project in West Java, I used a step selection function (SSF) framework, to understand how landscape structure affects the movement of Javan slow lorises within a fragmented mountain-agroforest landscape of Cipaganti, West Java. To investigate the movement and fine scale habitat selection of slow lorises I used one hour locations of 6 radio-collared slow lorises. The habitat and vegetation of observed and random steps was investigated in multiple variables such as presence of food trees and signs of human disturbance. For the analysis I paired observed steps (1h relocations) with 3 random habitat locations and used a conditional logistic regression to parameterize the SSF, which represents the probability of a focal slow loris to select a given step as a function of the habitat and vegetation factors surveyed. In average the slow lorises travelled about 450 m each night and most frequently they used a step length of about 0 – 50 m. My result reveals that slow lorises fine scale habitat selection is positively influenced by the presence of trees and tree trunk cover (indirect increasing the canopy cover and connectivity). They are also to a high extent positively affected by the presence of a feeding tree species, Calliandra calothyrsus. Surprisingly slow lorises selected steps associated with a higher number of fields (fields may indicate an increased biodiversity within the location). The results also indicate that slow lorises are limited in their movement by the presence of fields or rivers, which indicates that slow lorises are negatively influenced in their movement by a declining ability to move and forage within Cipaganti. I found no significant differences between sexes in their distance travelled. The recommendation for future conservation of slow lorises in Cipaganti is to prevent further habitat loss and fragmentation through activities that protect or maintain the present suitable slow loris habitat. Further research is needed to increase the knowledge of these primates’ abilities to live in this modified landscape.